A Biological-Realtime Neuromorphic System in 28 nm CMOS Using Low-Leakage Switched Capacitor Circuits

Mayr, Christian; Partzsch, Johannes; Noack, Marko; Hänzsche, Stefan; Scholze, Stefan; Höppner, Sebastian; Ellguth, Georg; Schüffny, René

doi:10.1109/tbcas.2014.2379294

Cited by 92 publications

(75 citation statements)

References 45 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…On the left are the mixed-signal designs (Schemmel et al, 2010; Benjamin et al, 2014; Park et al, 2014; Qiao et al, 2015; Mayr et al, 2016; Moradi et al, 2018), digital designs (Seo et al, 2011; Merolla et al, 2014; Davies et al, 2018; Frenkel et al, 2018) together with this work on are on the right. Toward large-scale spiking neuromorphic platforms, the key figures of merit are density, flexibility, synaptic plasticity, and energy consumed per SOP.…”

Section: Discussion and Future Workmentioning

confidence: 99%

Breaking Liebig’s Law: An Advanced Multipurpose Neuromorphic Engine

Wang

Schaik

2018

Front. Neurosci.

View full text Add to dashboard Cite

We present a massively-parallel scalable multi-purpose neuromorphic engine. All existing neuromorphic hardware systems suffer from Liebig’s law (that the performance of the system is limited by the component in shortest supply) as they have fixed numbers of dedicated neurons and synapses for specific types of plasticity. For any application, it is always the availability of one of these components that limits the size of the model, leaving the others unused. To overcome this problem, our engine adopts a unique novel architecture: an array of identical components, each of which can be configured as a leaky-integrate-and-fire (LIF) neuron, a learning-synapse, or an axon with trainable delay. Spike timing dependent plasticity (STDP) and spike timing dependent delay plasticity (STDDP) are the two supported learning rules. All the parameters are stored in the SRAMs such that runtime reconfiguration is supported. As a proof of concept, we have implemented a prototype system with 16 neural engines, each of which consists of 32768 (32k) components, yielding half a million components, on an entry level FPGA (Altera Cyclone V). We verified the prototype system with measurement results. To demonstrate that our neuromorphic engine is a high performance and scalable digital design, we implemented it using TSMC 28nm HPC technology. Place and route results using Cadence Innovus with a clock frequency of 2.5 GHz show that this engine achieves an excellent area efficiency of 1.68 μm2 per component: 256k (218) components in a silicon area of 650 μm × 680 μm (∼0.44 mm2, the utilization of the silicon area is 98.7%). The power consumption of this engine is 37 mW, yielding a power efficiency of 0.92 pJ per synaptic operation (SOP).

show abstract

Section: Discussion and Future Workmentioning

confidence: 99%

Breaking Liebig’s Law: An Advanced Multipurpose Neuromorphic Engine

Wang

Schaik

2018

Front. Neurosci.

View full text Add to dashboard Cite

show abstract

“…These restrictions put some doubt on whether complex learning mechanisms, as the one considered here, can be implemented exactly. Also, exact implementation of the synaptic sampling model seems infeasible on neuromorphic hardwares with configurable (but not programmable) plasticity, like ROLLS [64], ODIN [65] and TITAN [66] (see [67] and [68] for reviews). However, it might be possible to realize simplified, approximate, versions of synaptic sampling on these neuromorphic platforms.…”

Section: Comparison With Other Neuromorphic Platformsmentioning

confidence: 99%

Efficient Reward-Based Structural Plasticity on a SpiNNaker 2 Prototype

Yan

Kappel

Neumärker

et al. 2019

IEEE Trans. Biomed. Circuits Syst.

Self Cite

View full text Add to dashboard Cite

Advances in neuroscience uncover the mechanisms employed by the brain to efficiently solve complex learning tasks with very limited resources. However, the efficiency is often lost when one tries to port these findings to a silicon substrate, since brain-inspired algorithms often make extensive use of complex functions such as random number generators, that are expensive to compute on standard general purpose hardware. The prototype chip of the 2nd generation SpiNNaker system is designed to overcome this problem. Low-power ARM processors equipped with a random number generator and an exponential function accelerator enable the efficient execution of brain-inspired algorithms. We implement the recently introduced reward-based synaptic sampling model that employs structural plasticity to learn a function or task. The numerical simulation of the model requires to update the synapse variables in each time step including an explorative random term. To the best of our knowledge, this is the most complex synapse model implemented so far on the SpiNNaker system. By making efficient use of the hardware accelerators and numerical optimizations the computation time of one plasticity update is reduced by a factor of 2. This, combined with fitting the model into to the local SRAM, leads to 62% energy reduction compared to the case without accelerators and the use of external DRAM. The model implementation is integrated into the SpiNNaker software framework allowing for scalability onto larger systems. The hardware-software system presented in this work paves the way for power-efficient mobile and biomedical applications with biologically plausible brain-inspired algorithms.

show abstract

“…This point, in turn, creates an opportunity to design and develop the complex neuromorphic circuits with advance computing ability. Considering the fact that the design of the basic units of the nervous system in the form of hardware in neuromorphic systems has attracted a lot of attention, 3 in recent years, systems efficiency and reliability have increased by inspired hardware from brain in the form of analog [4][5][6] and digital 7-10 neuromorphic architecture. In addition, recent studies have focused on neuron-astrocyte interactions and synaptic plasticity [11][12][13][14] and have designed neuromorphic architecture based on neural cells interactions.…”

Section: Introductionmentioning

confidence: 99%

Digital realization of the proposed linear model of the Hodgkin‐Huxley neuron

Amiri

Nazari

Faez

2019

Circuit Theory & Apps

View full text Add to dashboard Cite

Summary It seems that computing systems that imitate the brain can be achieved by integrating of electronics and neuroscience. In recent years, neuromorphic systems have been developed by the fusion of electronics and neuroscience. Since neurons are the basis of neural systems, constructing the optimized digital neuron plays a critical role in neuromorphic applications. Furthermore, the dynamic of ionic channels, which are the causative agents of synaptic plasticity is the main characteristic of biological neurons. The Hodgkin‐Huxley neuron model is a mathematical description of biological neuron that is widely used in neuroscience to explore the relation of action potential propagation and information transmission. This model consists of nonlinear differential equations, which approximates the electrical characteristics of excitable cells such as neurons and cardiac muscle. In this paper, a simplified version of the Hodgkin‐Huxley neuron model was proposed by substituting its complex nonlinear dynamics with linear ones. Next, a digital circuit for the proposed linear model was designed, which can be implemented on a low‐cost hardware platform, such as field‐programmable gate array (FPGA). Comparison of MATLAB simulation of original model and Vivado simulation of the proposed digital circuit confirm that obtained results are in good agreement. The proposed digital circuit compared with the earlier circuits is more successful in terms of replicate essential characteristics of spiking responses and ionic currents in the biologically plausible model of neural activities. This new digital design will be applicable in developing neuro‐inspired chips and exploring the brain's functionality in information processing.

show abstract

A Biological-Realtime Neuromorphic System in 28 nm CMOS Using Low-Leakage Switched Capacitor Circuits

Cited by 92 publications

References 45 publications

Breaking Liebig’s Law: An Advanced Multipurpose Neuromorphic Engine

Breaking Liebig’s Law: An Advanced Multipurpose Neuromorphic Engine

Efficient Reward-Based Structural Plasticity on a SpiNNaker 2 Prototype

Digital realization of the proposed linear model of the Hodgkin‐Huxley neuron

Contact Info

Product

Resources

About